If you run a workflow from the Workflows page, you will need to select the correct "root entity type" from the dropdown. This article gives guidance for determining the root entity type when you have several tables to choose from.
Root entity type: Overview
The root entity type is the smallest piece of data a workflow can use as input. Selecting from the "root entity type" dropdown (Step 1 in the workflow configuration form) tells the workflow which table has the input metadata needed to run the workflow.
Note that tables can be nested. For example, if your workflow runs on sets of data (arrays), the root entity type could be a sample_set or specimen_set table, even though links to the data files are ultimately in a sample or specimens table.
Finding the root entity type If you can run your workflow on a single entity (like a specimen or sample)
The root entity type is that entity (i.e. specimen
or sample
)
If your workflow takes an array as input and cannot run on a single file
1. The root entity type could be a set table (i.e. sample_set
or specimen_set
).
2. The root entity type could be an entity table with a column of arrays.
If you're running a somatic workflow (on tumor/normal pairs)
The root entity type is pair
Root entity type examples
See the examples below for ways to determine the input entity type your WDL expects (could be a single entity table, a set table, or a pair table).
-
If your workflow can run on a single entity, the root entity will be a table of single entities (i.e.
sample
orspecimen
).How to verify
Look in the Inputs tab of the workflow form. The Input Type will be "File". -
If your workflow runs on an array of entities, the root entity will be a set table (i.e.
sample_set
orspecimen_set
).How to verify
Look in the Inputs tab of the workflow form. The Input Type will be "Array[File]". -
If you are running a somatic workflow, the root entity will be
pair
. The file type could be File or Array[File] (depending on the workflow).